In honor of Independence Day, ASNT Pulse is republishing this article on nondestructive testing (NDT) conducted on the Liberty Bell. It originally appeared in the NDT Solution column, edited by G.P. Singh, from the September 2003 issue of ASNT’s Materials Evaluation. It appears here with minor edits.
___
This article demonstrates the application of NDT methods in preserving one of the symbols of our national heritage. The author demonstrates the unique applicability of electrical conductivity as an NDT method in assessing the surface damage and overall condition of the Liberty Bell. Readers should find this article very informative and interesting.
G.P. Singh
Associate Technical Editor
___
By Louis R. Truckley
.
Introduction
In April 2001, the Liberty Bell was vandalized by a visitor wielding a sledgehammer. Due to this attack, and in preparation to relocate the bell to its new home in Liberty Bell Center, NDT was performed to assess the hammer damage and to determine its overall condition. These tests employed radiographic, liquid penetrant, eddy current, ultrasonic, and electrical conductivity testing. The tests were requested by the National Park Service, custodian of the Liberty Bell, and the Philadelphia Museum of Art, which serves as a technical consultant to the Park Service for the conservation of the bell.
This paper discusses the electrical conductivity tests performed on the Liberty Bell in May 2001. The purpose of these tests was to examine the homogeneity of the bronze alloy material used for the large bell casting and to produce electrical conductivity maps of the surface of the approximately 943.5 kg (2080 lb) casting. Electrical conductivity testing is an NDT methodology that measures a metal’s ability to allow the flow of electrons through its lattice structure. The unit of measure for conductivity testing is commonly expressed as a percentage based on the International Annealed Copper Standard (IACS).
Performing modern NDT on this nationally honored symbol was an interesting and rewarding experience. It was an extraordinary opportunity to examine the same bell that tolled to announce the first reading of the Declaration of Independence during the lives of the founders of our nation.
Purpose
The Liberty Bell is a bronze alloy casting cast by two Philadelphia foundry men, John Pass and John Stow, in 1753 (Franklin Institute, 1962). Bronze alloys have been used in the casting process since prior to 700 BC. The particular bronze alloy used for the Liberty Bell is referred to as “bell bronze.” Bell bronze may also be called “tin bronze” because of the high tin content in this particular alloy. To this day, bell bronze is the preferred material for bell casting because it yields a distinct tone that other metals cannot match. Bell bronze is considered a hard material due to its 20% or greater tin content.
The casting alloy used for the Liberty Bell actually contains 10 different elements. In 1975, an X-ray fluorescence analysis was performed on the bell by representatives from Winterthur Museum and the DuPont Company by sampling 10 separate points around the rim of the bell. The varying metal contents revealed in this test are shown in Table 1 (Kimball, 1997).
In fact, it was this varying alloy content that generated the technical interest in performing an electrical conductivity test on the bell. As alloying elements are added or varied within a base material, the physical property of electrical conductivity can change to a large degree. What added to this interest in performing an electrical conductivity test was the fact that, in 1753, the Pass and Stow foundry equipment was not capable of melting and pouring large quantities of bronze in one melt or one heat. Therefore, the bronze material had to be melted and poured in multiple small batches to cast such a large structure as the Liberty Bell. These small batches used throughout the casting process could vary the metallurgical structure between melts, which could significantly change the conductivity values across the surface of the bell. With this interest in the homogeneity of the bronze casting, an electrical conductivity test was planned and employed on the bell on 6 May 2001.
Surface Layout
In order to produce an electrical conductivity map of the Liberty Bell, it was necessary to provide a grid on the surface to relate particular measurements to specific locations. The Liberty Bell is quite a large structure, with a surface area exceeding 2.8 m2 (30 ft2). The thickness of the bell’s wall is 76 mm (3 in.) near the rim and tapers down to approximately 32 mm (1.3 in.) at the crown. The overall dimensions are identified in Figure 1. Additionally, Figure 1 shows the bell’s “canons,” which protrude from the top and support its weight.
.
It was desired to obtain electrical conductivity measurements on a grid spacing of 150 to 200 mm (6 to 8 in.). A measurement grid was produced by placing a series of 22 mm (0.9 in.) diameter white dot stickers on the surface of the bell and canons as shown in Figure 2.
.
.
As a precaution to protect the bell’s fragile patina and its noble appearance, a layer of masking tape was trimmed to fit beneath the white dot stickers. The white dot rows were then identified in a vertical direction starting at the crown with row 1 and ending at the rim with row 7. Then, beginning with number 1 at the repaired crack location and extending in a clockwise direction, each individual white dot was identified with a number. The flat sides of the canons were also identified with white dots. These eight dots began at the position closest to the repaired crack location and extended in a clockwise direction.
Digital pictures were then taken at various angles to capture the entire measurement grid. These pictures were later used to create a graphic map by matching a conductivity measurement with each white dot shown.
Equipment Standardization and Measurement
For this test, a portable direct reading electrical conductivity instrument was used with a commercially available set of conductivity standards. The electrical conductivity instrument and the reference standards were certified for calibration. A sufficient period of time was allowed to ensure that the test instrument, reference standards, and the bell were at the same temperature. The following test variables that could affect accurate readings were considered prior to instrument standardization: coatings/cladding, thin gauge, and a convex test surface (Boeing, 1997). No coatings, other than the patina, were present on the test surface of the Liberty Bell. The thin gauge variable was not an issue since the thinnest section of the bell exceeded 25 mm (1 in).
Finally, the convex test surface was not an issue since the smallest radius of curvature exceeded 310 mm (12.2 in.). It was also understood that the presence of casting discontinuities would lower the readings; however, no correction factors could be applied for this phenomenon. With these variables considered, the electrical conductivity instrument was standardized using the following four conductivity reference standards:
- 1.068% IACS (0.619 MS/m)
- 3.514% IACS (2.038 MS/m)
- 8.184% IACS (4.747 MS/m)
- 29.61% IACS (17.17 MS/m)
During the equipment standardization process, the instrument was adjusted to repeatedly display the value of each reference standard to within 0.5% IACS (0.3 MS/m). Once this standardization was achieved, the measurement process on the Liberty Bell began with the eight readings on the canons. Then, row 1 was measured in a clockwise direction, followed by row 2 and so forth. A total of seven rows of readings were collected in addition to the eight measurements on the canons.
When this measurement process was completed, a total of 113 individual electrical conductivity values were recorded with respect to their position on the bell using the white dots as visual reference indicators. The equipment standardization was verified with the reference standards at approximate 15 min intervals during the measurement process and finally at the completion of the measurement process.
Data Interpretation
The electrical conductivity measurements were then incorporated into a digital image in order to map the bell’s conductivity values. Four of these maps were prepared to display each quadrant of the bell. Figure 3 shows one of these maps as an example. On each map, the conductivity values are shown as a percentage IACS. Additionally, the average reading, standard deviation, minimum reading, maximum reading and range are noted for each quadrant map. The primary objective during this phase of data interpretation was identifying the average overall conductivity value and determining if there were any regions exhibiting significant material property changes.
.
.
The average conductivity value for the Liberty Bell is 3.9% IACS (2.3 MS/m). The readings were remarkably consistent over the entire surface of the bell. In a few locations, a change of 1.5 to 2% IACS (0.87 to 1.2 MS/m) was measured; however, these areas exhibited noticeable indications of casting discontinuities such as porosity and gas holes open to the surface. In fact, when adjacent areas were remeasured, the conductivity values were typically in the range of 3.6 to 4.4% IACS (2.1 to 2.6 MS/m) as was the case with most of the bell.
The plan in establishing this measurement grid was to locate zones of large variation and then tighten the grid in these regions to further assess major material changes; however, no areas of material changes that would affect electrical conductivity were detected.
The next concern to be addressed was determining the appropriate electrical conductivity value for a copper alloy cast in 1753 with approximately 70% copper, 25% tin, 3% lead, and trace amounts of seven other elements. The conductivity of the copper alone would be near 100% IACS (58 MS/m); however, the electrical conductivity of copper is significantly lowered with the presence of small quantities of alloying elements and impurities.
One approach to determine the appropriate conductivity value was to refer to the Metals Handbook—Properties and Selection: Nonferrous Alloys and Pure Metals (American Society for Metals, 1979) for properties of cast copper alloys. This revealed that common bronze alloys have an electrical conductivity value between 10% IACS (5.8 MS/m) and 12% IACS (7 MS/m); however, the tin content in the listed “tin bronze” alloys was much lower than the bell’s alloy (less than 12%) and the copper content was much higher than the bell’s alloy (87 to 89%).
A strategy was considered to create a chart depicting conductivity versus alloy for the tin bronze alloys listed in the AMS handbook and to extrapolate the electrical conductivity of the bell by indicating that the conductivity values will decrease as alloying elements increase; however, a sufficient quantity of data points to produce such a curve was not available.
Another approach to determine the appropriate conductivity value for the bell was to contact bell manufacturers and inquire if electrical conductivity data are available. A bell manufacturer was contacted in the state of Maryland to discuss conductivity data or possibly obtain a sample of bell bronze for testing; however, this effort did not produce any useful data. A second bell manufacturer was also contacted to no avail.
The final approach to determine the appropriate conductivity value for the Liberty Bell was to test a bell fabricated from a similar casting process and bell bronze alloy. Christ Church in Philadelphia, Pennsylvania, has been a standing place of worship since the 1720s. The bell tower there contains bells that were cast in 1754 by Whitechapel Bell Foundry using a similar process and bronze alloy as the Liberty Bell. These bells are still in use today (Christ Church Preservation Trust, 2001). Figure 4 contains pictures of the 61 m (200 ft) high steeple of Christ Church and the series of bronze bells that are suspended in a carriage in the steeple.
.
.
A climb into the bell tower with equipment and standards in hand resulted in a conductivity measurement of two of these 1754 bells, identified as numbers 4 and 6, as shown in Figure 5. A sampling of 14 conductivity readings were obtained from the number 4 bell and 12 from the number 6 bell. These readings were sampled from top to bottom and on both sides of the bells (180° apart).
.
.
It was encouraging to find that the electrical conductivity results of the 1754 Christ Church bells matched the readings obtained from the Liberty Bell. The number 4 bell average conductivity value is 3.7% IACS (2.1 MS/m) and the number 6 average conductivity value is 3.8% IACS (2.2 MS/m). As stated earlier, the average conductivity value for the Liberty Bell is 3.9% IACS (2.3 MS/m).
Conclusions
This test yielded two primary conclusions. The first is that the Liberty Bell exhibits minor variation in material properties that affect electrical conductivity. The second conclusion is that the electrical conductivity values of the Liberty Bell are consistent with other bells functioning since 1754 and cast using a similar process and bronze alloy.
Acknowledgments
The author would like to acknowledge and thank the National Park Service, Independence National Historic Park, especially Karie Diethorn, curator of the bell, for the opportunity to test the Liberty Bell; the staff of Christ Church, Philadelphia; Andrew Lins, chief conservator of the Philadelphia Museum of Art; and Ira Sherman and the Boeing Company for providing pro bono resources.
References
- ASM, 1997, Metals Handbook—Properties and Selection: Nonferrous Alloys and Pure Metals, ninth edition, Vol. 2, Materials Park, Ohio, American Society for Metals
- Boeing Company, 1997, Boeing Process Specification, BAC 5651 Eddy Current Electrical Conductivity Inspection, Revision A, Seattle, Washington, Boeing Company
- The Christ Church Preservation Trust, 2001, The Organ and Bells of Christ Church, Philadelphia, Pennsylvania
- The Franklin Institute, 1962, Report of the Committee for the Preservation of the Liberty Bell, Philadelphia, Pennsylvania
- Kimball, David, 1997, The Story of the Liberty Bell, New York, Eastern National, 1997.
.
_______
.
- Louis R. Truckley wrote this article for the NDT Solution column in ASNT’s Materials Evaluation in 2003.
- G.P. Singh was an Associate Technical Editor for Materials Evaluation.
To read more about NDT performed on historical objects, visit the NDT Library.